Skip to content

Use Cases

Fraud prevention

Identity fraud is becoming increasingly prevalent in the digital world. The security of companies and their users needs to adopt measures to detect these attacks and threats.

A common type of fraud is the attempt to register multiple times with different identities for a given service (i.e. opening banking accounts, insurance policies, etc.). In these frauudlent cases, it is common to provide fake documentation in which the data has been manipulated (name, surname, etc.) but not the picture as that would be detected by the staff in charge of the onboarding (a.k.a subject registration). By simply taking a picture of such documentation, and thanks to dasFaceBond, that picture could be used to check whether or not the database of subjects already includes a subject with the same face regardless of the data provided in the past or the present. Getting this confirmation in a matter of seconds becomes critical to avoid providing the mentioned service to the potential fraudster. Lastly, raising the alert of a potential fraud attempt would allow the user to put the necessary additionally measurements to prevent the subject to perform new fraudulent activities in the future.

  1. Forensic search of subjects based on photographs on the database of the solution.

  1. Search for duplicate identities in a database of facial images. This makes it possible to identify people with different identities who are really the same person.

Further details of usage are available in the user's guide.

Deduplication

Another use case in which a 1:N type of comparison (1:N stands for a one voice sample to be compared with N voiceprints) and the so-called 'clusterization' (to group the current records of a database by subject) allows to detect of duplicated entries in the database. This could be very useful for instance when there are legacy databases that need to be merged (for instance in a fusion of two Banking entities) and having the same subjects registered in both databases is something that may arise. Ideally, in the merged database, each subject should appear only once with the historical data of each database merged into one single register, but this may turn out to be difficult sometimes if the metadata of the register is not identical. Thanks to clusterization, duplicate entries can be easily found, allowing to implementation of a mechanism to detect these cases and to put in place adequate procedures.